In part one I said the data can be moved out of Printavo, all of it, if you do it in the right order. When I shared my export checklist in a Printavo users group, some people doubted that a full export and import was really possible.
That is a fair thing to doubt, so here is the result. And here is the part I left out of part one: getting the data out was the easy part. Making it usable took another month.
One customer, back to 2019
This is one of my longest-running customers. The list scrolls from this year back to their first order on record, in May 2019. Then it opens that order with its line items, and a 2020 order with the mockups still attached.
Their name, contacts and city are under solid bars, and so is every dollar amount, theirs and mine. I used bars instead of blur because blurred text can sometimes be read back. I recorded it on the test copy of my system rather than the live one, because my recording tools are not allowed to run against the system the shop works in.
Across the whole shop, when I audited the import in August, order totals matched Printavo on 5,776 of 5,800 orders. Today 8,011 old mockups sit on the 3,876 old orders they belong to.
Why I brought all of it
Starting fresh is tempting. Export a customer list, keep the old PDFs in a folder, and begin again.
Here is why I did not. Since we switched in July, about four out of every five jobs we have booked came from a customer who was already in the old system. The typical one had ordered from us about five times before, and about a third of them go back to 2021 or earlier.
Nobody ever comes back years later and says, I’d like to order that thing you quoted me. People want to reorder what they actually bought: the same shirt, the same color, the same print in the same spot. That history is the business.
The old price matters too. Sometimes I requote an old job at what feels like enough money, only to realize that’s why I charged double for it years ago.
What went wrong after the import
The records themselves came over fine. Then the problems started, and almost every one had the same shape: the data was there, and nobody could use it.
- Old orders opened empty. About three weeks after we went live, we found that 6,401 old orders opened with nothing on screen. The line items had all come over, but not one of the imported ones was attached to anything the screen actually draws. My team had worried about losing the ability to look up old jobs, and they were right. We rebuilt the old orders on August 10, and I canceled Printavo four days later.
- Search could not find old jobs by name. The import read every job’s name from Printavo and then saved it nowhere, so until we fixed it on August 6, search was blind to old job names. Today about 3,700 old orders can be found by name.
- Hundreds of finished jobs landed on the production board. Old jobs that were done but never closed out in Printavo showed up as live work on day one, 782 of them. Archiving them took the board from 834 jobs to 52.
- Old quotes came over as if they were orders. 625 quotes that customers never approved came in looking like real, unpaid jobs. Before we deleted them on August 17, we checked for evidence of real work, and that check caught two paid jobs still sitting in a quote status. They stayed. A status tells you how your shop used Printavo, not whether a job was done or paid.
What did not come over
This is the honest list.
- Message threads and payment detail. All of it is saved in cold storage, including 3,200 message threads and the payment history, but none of it lives in the new system. When someone needs an old message, they ask me and I pull it from the archive.
- What the shirts cost us. The old orders carry what we charged, not what the garments cost us, so the margin on an old job is unknowable.
- Some ink and placement notes. On 527 old print locations, notes like “Left Chest / PMS ORANGE 130 C” came over but do not show on any screen yet. That fix is still open.
- 29 orders. A handful of Printavo orders never came over at all, 29 of them. They are in the archive, and adding them is a someday job.
- Anything typed freehand stays freehand. In a sample of our earliest 2019 orders, 80% of garment lines had been typed in by hand instead of picked from Printavo’s catalog. An import cannot turn a typed line back into a catalog product.
If you are planning your own move
Part one covered what to export and when. This is what I would check after the import, while the old account is still open to compare against:
- Open a dozen old orders on the screen your team actually uses. A record that exists in the database but shows nothing on screen is not there.
- Search for old jobs by name, by customer and by garment.
- Look at your production board before your team does.
- Bring over real work, not every quote.
- Before deleting anything by status, check for payments and files. Your statuses mean whatever your shop used them for.
- Keep imported history away from anything that touches money: customer balances, payment reminders and your sales tax report.
- If you switch mid-month, pull that month’s sales tax report from the old system before you cancel. We missed this one.
- If a question can only be answered from inside the old account, answer it before you cancel. After that, nobody can check.
Was it worth it
Yes. The relief I wrote about in part one has held. But the export was not the cost. The month after it was, and because the system is one I built for my own floor, that month was on me.
If you are moving to a system someone else sells, ask them to import a sample of your real data first, then open your oldest customer’s oldest order on screen. Before you sign, not after.
The one-page version
The export order from part one is a free one-page checklist, with the exact menu paths for the reports the API will never give you.
Get the checklist: what to export before you cancel Printavo